CANVAS-BACK 125 



winter on the lakes of the San Pedro Martir Mountains, though it is not seen on the coasts of Lower 

 California. The southernmost record is that of specimens taken at Dueflas (Salvin „ . . 

 and Godman, 1897-1904) and at Swan Isle, Guatemala (U.S. Biological Survey). 



Migration 



The migration area and dates for the Canvas-back are so similar to those of the Red-head that great 

 detail would be superfluous. Suffice it to say that in the northwestern States the species is known 

 chiefly as a bird of passage. In the autumn many migrate eastward along the Great Lakes, though 

 they do not go so far east as does the Red-head. It is or used to be a much rarer bird than the latter 

 even in Ontario, while in the New England States and in the Maritime Provinces it can hardly be 

 considered as anything but accidental. The spring migration probably follows the Mississippi Valley 

 more than the Great Lakes route. 



The data as to the migration are far too meager to permit of any conclusions. Representative 

 dates, such as those given by Cooke (1906), are misleading in so far as they give rise to the impression 

 that the flocks pass north or south en masse. This is not the case, for the migration is on the whole 

 spasmodic and so dependent on local weather conditions that only great numbers of dates would per- 

 mit of general conclusions. The fact that Canvas-backs appear in the northern States at the time 

 when others are still in Mexico or Texas and that in autumn many are in their winter homes in the 

 South before others begin to appear on the North Pacific coast gives some idea of how difficult it is to 

 work out a logical account of the seasonal movements. In view of the inadequacy of our knowledge 

 one can only give a few dates as a foundation for further detailed work. In spring many Canvas-backs 

 leave the Gulf Coast in February and some appear in Kansas, Missouri, and Indiana by the end of 

 that month. But the bulk pass the north-central States during March or the first half of April and do 

 not reach the southern parts of Canada till the second half of April. On Lake Athabasca in 1920 they 

 did not appear before May 12. On the Pacific Coast they leave California in March and April, and 

 pass over Puget Sound in the latter month. Sanford (Sanford, Bishop and Van Dyke, 1903) mentions 

 seeing some as far south as Tampico in late April. 



The females and young are the first to start south in the autumn. Some regions seem to attract the 

 birds of a different age (and sex?). Leopold (1921) noticed that all the Canvas-backs in the Albu- 

 querque region of New Mexico were small and immature, while those from Lake Burford were large 

 and fully feathered even early in the season. There seems to be a great excess of males among those 

 Canvas-backs that winter on the lakes of central New York, as many as four or six males to one 

 female. 



In the autumn some of the Canvas-backs pass south into the United States as early as late Sep- 

 tember, but the great mass appear in our northern States during mid-October, sometimes delayed 

 until the last of the month, and reach their wintering quarters on the Atlantic seaboard toward the 

 end of the same month or the beginning of November. They keep increasing in numbers until De- 

 cember. In Mexico they hardly appear before November, while on the Pacific Coast the great bulk 

 comes in during November and December. 



GENERAL HABITS 



The Canvas-back is our most famous duck and on the whole deserves its reputation, 

 if for nothing else than because it has won out in a terrific contest with man and his 

 most artful devices. This ought to be fame enough for any one bird. 



Not content with leaving it among its relatives, recent writers have complimented 

 this duck by giving it generic rank, forgetting, I think, that a genus is meant to convey 

 the idea of relationships, just as much as it is to accentuate differences. In fact, one 



